Many of today's fiber optic communication devices use lasers as a light source for digital communication. Additionally, other devices, such as optical mouse input devices or laser pointers also use lasers in their respective operation.
These lasers generate optical signals for digital fiber-optic transmissions and can sometimes generate light powerful enough to damage the human eye. For example, a typical fiber-optic communication device includes a connector for connecting a fiber-optic cable in optical alignment with the laser. Therefore, if a person looks into the end of the optical fiber, then the laser beam may enter his eye. Alternatively, if a person disconnects the cable and looks directly into the connector, the laser beam may also enter his eye. Furthermore, broken or cut fiber-optic cables may also allow a laser beam to inadvertently enter a person's eye.
If the average power of the laser beam is high enough, then the beam may damage a person's retina. Consequently, in situations where the laser beam may enter one's eye, safety is of a concern. As a result, to conform to eye-safety standards fiber-optic communication devices typically include laser fault-detection and disable/enable circuitry.
One example of a fiber-optic communication device with a conventional disable circuit is shown in FIG. 1. A transceiver module 100 includes a receiver 101, a transmitter 102, and a controller/control circuitry 103 and is typically part of a conventional fiber-optic communication system in which the transceiver module 100 is coupled to a host 105. The receiver 101, transmitter 102, and controller 103 are coupled to the host 105 via a communication link 106 and the receiver 101 and transmitter 102 are coupled to the controller 103 via respective communication links 107 and 108. The operation of a conventional transceiver module such as module 100 is known in the art and is described in greater detail in U.S. patent application Ser. No. 10/327,216 filed on Dec. 20, 2002, and which is assigned to Agilent Technologies of Palo Alto, Calif.
The transmitter 102 includes at least one laser driver 105 coupled to a laser 120 and a disable circuit 110. The transmitter 102 further includes fault-detection circuitry 130 for detecting faults in the transmitter 102. Upon detection of a fault, the fault-detection circuitry 130 sends a signal via fault flag signal line 150 for disabling the laser driver 106, and sends a fault signal on a fault signal line 151 to the host 105. By convention, the host 105 typically generates a disable signal on a disable line 152 to the disable circuit 110 which is, in turn, coupled to the laser driver 106. The disable circuit 110 generates a shutdown signal on a shutdown-signal line 153 that also disables the laser driver 105 and, thus, the laser 120 cannot be turned on. The disable circuit 110 generates the shutdown signal for a number of reasons, including safe laser power levels being exceeded and a disable signal from the host 105.
A problem arises, however, in specific situations when the disable line 152 to the disable circuit 110 is cycled (repetitive alternating signals) at a speed close to the fault response time (described below) of the fault detection circuitry 130. In particular, when an unrecoverable fault in the transmitter 102 has occurred, a typical host 105 will cycle the disable signal on the disable line 152 to the faulted transceiver module 100 to reset the module to its default power-up state or some other initial state. Thus, it is possible for a transceiver module 100 to get into a state where the laser is turned on and off repeatedly. For example, fault detection circuitry 130 in the transceiver module 100 may detect a fault situation, such as the laser 120 is operating at an excessive power level, and, as a result, send a fault signal to the host 105. The host 105, in response to the fault signal sends a disable signal back to the transceiver module 100, which in turn causes the disable circuit to generate a shutdown signal to disable the laser 120 and reset the fault detection circuitry 130. The host 105 then attempts to restart the transmitter 102 by sending a re-enable signal to the transmitter 102 through the disable line 152. Upon receiving the re-enable signal, the disable circuit 110 terminates the shutdown signal, and the laser 120 will turn back on again. If there is still a fault present, the laser 120 is turned off once again because the fault detection circuitry 130 still detects a fault. This is a common method used to clear a transmitter fault that may be transient and recoverable. Thus, if the fault is not recoverable and this re-enable/disable cycle is repeated at a speed close to the fault response time, the total average power transmitted by the laser may still exceed eye-safe power levels. This phenomenon is further discussed below in conjunction with FIG. 2.
FIG. 2 is a timing diagram of signals in the transceiver module 100 of FIG. 1 in the above-described situation when a conventional disable circuit allows a laser to exceed eye-safe power levels.
During a period 210 of normal operation, i.e., a fault 201 condition is not present, the laser operating power 204 (LOP) has a normal average power 211, which meets eye-safety standards.
When a recoverable fault 220 occurs, the fault detection circuitry 130 (FIG. 1) detects the fault 220, shuts off the laser by generating a laser off signal 203, and sets a fault flag 205 at point 224 that the host 105 (FIG. 1) reads either on a polling or an interrupt basis. A finite amount of time exists between the occurrence of the fault 220 and the assertion of the fault flag 205 and laser off 203 signals which is called the fault response time 250. In response to the fault flag 205, the host 105 asserts a disable signal 202 (shown as active high) at point 221 which is some time after the assertion of the fault flag signal 205. Typically, the laser will have an LOP 204 that is an excessively high value 251 during the fault response time 250 because a fault 201 has occurred and the laser off signal 203 has not been asserted yet. During this time, the laser 120 is still generating a higher level 251 of average power than allowed, although not for a time period long enough to cause damage to one's eye. The disable signal 202 clears the fault flag signal 205 at point 226 but still prevents the laser 120 from being turned back on as long as the disable signal 202 is asserted. While the laser 120 is off, the LOP 204 is zero as shown at 223. Typically, the host 105 keeps the disable signal 202 active for a predetermined time that is deemed sufficient for the fault 220 to be corrected. Then, in response to the deassertion of the disable signal 202, the laser off signal 203 is deasserted at point 227, which will allow the laser 120 to be turned back on for normal operation. If the fault is corrected, then the transceiver module 100 continues to operate normally such that the LOP 204 has an average-power level that is less than or equal to the maximum eye-safe level.
The problematic situation described above occurs when an unrecoverable fault 230 occurs. In some instances, the unrecoverable fault 230 cannot be corrected through automatic means (cycling the disable line 152 or power), and thus, may hold the average power level of the LOP 204 to an unsafe level 240. As in a normal situation, the unrecoverable fault 230 is detected by the detection circuitry 130, the fault flag 205 is set at point 235 after the fault response time 250 which causes the laser-off signal 203 to be asserted, also at point 235, thereby turning off the laser. In response to the fault flag 205, the host 105 asserts a disable signal 202 at point 231 to clear the fault flag 205. After the given amount of time determined by the host 105, the disable signal is deasserted at point 238, which in turn, allows the laser-off signal 203 to deassert. Different from the normal situation, however, the unrecoverable fault 230 has not cleared and the LOP 204 may go to a maximum (non-eye-safe) value at point 233 after the laser-off signal 203 is deasserted but before the fault flag 205 is set again. The fault detection circuitry 130 will detect the fault 230 again and assert the fault flag 205 and laser off 203 signals at point 240, but the LOP 204 will remain at maximum power during the fault response time 250 and the process could repeat indefinitely. Thus, the overall average-power level 240 in the unrecoverable fault 230 situation can exceed the maximum eye-safe threshold if the duty cycle of the laser is large enough.
One solution to this problem is to count the number of assertions of the disable signal 202 within a predetermined time period and to permanently disable the laser 120 if the number of assertions is greater than a predetermined threshold. For example, one could set the controller 103 to permanently (until the module 100 is reset from an external source, typically by a power cycle) disable the laser 120 if the disable signal 202 is asserted more than 3 times within a 50 millisecond period.
A problem with this solution, however, is that once the disable assertion threshold is exceeded, the transceiver module 100 requires a power cycle to be enabled again. That is, the module 100 must be turned off and then on again to clear the fault.